eks-agent: allow endpoint override when querying pre-existing cluster info

[etungsten]

Issue number:
Partially addresses #565. Only for EKS service endpoint.

Description of changes:

    agents: add endpoint override support in eks provider
    
    We add a configuration in EKS cluster resource agent to allow EKS
    service endpoint override when querying cluster information for a
    pre-existing cluster.

    eks-resource-agent: use AWS API calls to gather created cluster info
    
    Instead of calling eksctl for certain cluster info, use AWS API calls
    and AWS-CLI calls to gather cluster metadata.

Testing done:

Normal pre-existing cluster created using eksctl:
Deployed a manifest for the cluster, EC2 instances, quick sonobuoy test.

---
apiVersion: testsys.system/v1
kind: Resource
metadata:
  name: ipv6-bottlerocket
  namespace: testsys
spec:
  agent:
    timeout: 70d
    name: eks-provider
    image: eks-resource-agent:latest
    keepRunning: true
    configuration:
      creationPolicy: never
      clusterName: ipv6-bottlerocket
      region: us-west-2
  dependsOn: []
  destructionPolicy: never
---
---
apiVersion: testsys.system/v1
kind: Resource
metadata:
  name: ipv6-bottlerocket-instances
  namespace: testsys
spec:
  agent:
    timeout: 70d
    name: ec2-provider
    image: ec2-resource-agent:latest
    keepRunning: true
    configuration:
      clusterName: ${ipv6-bottlerocket.clusterName}
      clusterType: eks
      instanceCount: 2
      instanceProfileArn: ${ipv6-bottlerocket.iamInstanceProfileArn}
      nodeAmi: ami-
      region: us-west-2
      subnetIds: ${ipv6-bottlerocket.publicSubnetIds}
      instanceTypes: ["m5.large"]
      endpoint: ${ipv6-bottlerocket.endpoint}
      certificate: ${ipv6-bottlerocket.certificate}
      clusterDnsIp: ${ipv6-bottlerocket.clusterDnsIp}
      securityGroups: ${ipv6-bottlerocket.securityGroups}
  dependsOn: [ipv6-bottlerocket]
  destructionPolicy: onDeletion
---
apiVersion: testsys.system/v1
kind: Test
metadata:
  name: ipv6-bottlerocket-test
  namespace: testsys
spec:
  agent:
    timeout: 70d
    name: sonobuoy-test-agent
    image: sonobuoy-test-agent:latest
    keepRunning: true
    configuration:
      region: ${ipv6-bottlerocket.region}
      kubeconfigBase64: ${ipv6-bottlerocket.encodedKubeconfig}
      plugin: "e2e"
      mode: quick
  dependsOn: []
  resources: [ipv6-bottlerocket-instances, ipv6-bottlerocket]

Resource gets created, test runs to completion:

$ cli --kubeconfig etung-testsys.kubeconfig status
 NAME                                            TYPE                         STATE                                           PASSED                     FAILED                    SKIPPED 
 ipv6-bottlerocket-test                          Test                         passed                                               5                          0                       7206 
 ipv6-bottlerocket                               Resource                     completed                                                                                                    
 ipv6-bottlerocket-instances                     Resource                     completed     

For a pre-existing cluster in a different EKS service endpoint:

Deployed the following manifest:

---
apiVersion: testsys.system/v1
kind: Resource
metadata:
  name: beta-x86-64-aws-k8s-128
  namespace: testsys
spec:
  agent:
    timeout: 7d
    name: eks-provider
    image: eks-resource-agent:what
    keepRunning: true
    configuration:
      eksServiceEndpoint: https://api.beta.us-west-2.wesley.amazonaws.com
      creationPolicy: never
      clusterName: beta-x86-64-aws-k8s-128
      region: us-west-2
  dependsOn: []
  destructionPolicy: never
---
apiVersion: testsys.system/v1
kind: Resource
metadata:
  name: beta-x86-64-aws-k8s-128-instances
  namespace: testsys
spec:
  agent:
    timeout: 7d
    name: ec2-provider
    image: ec2-resource-agent:latest
    keepRunning: true
    configuration:
      clusterName: ${beta-x86-64-aws-k8s-128.clusterName}
      clusterType: eks
      instanceCount: 2
      instanceProfileArn: ${beta-x86-64-aws-k8s-128.iamInstanceProfileArn}
      nodeAmi: ami-
      region: us-west-2
      subnetIds: ${beta-x86-64-aws-k8s-128.publicSubnetIds}
      instanceTypes: ["m5.large"]
      endpoint: ${beta-x86-64-aws-k8s-128.endpoint}
      certificate: ${beta-x86-64-aws-k8s-128.certificate}
      clusterDnsIp: ${beta-x86-64-aws-k8s-128.clusterDnsIp}
      securityGroups: ${beta-x86-64-aws-k8s-128.securityGroups}
  dependsOn: [beta-x86-64-aws-k8s-128]
  destructionPolicy: onDeletion
---
apiVersion: testsys.system/v1
kind: Test
metadata:
  name: beta-x86-64-aws-k8s-128-test
  namespace: testsys
spec:
  agent:
    timeout: 7d
    name: sonobuoy-test-agent
    image: sonobuoy-test-agent:latest
    keepRunning: true
    configuration:
      region: ${beta-x86-64-aws-k8s-128.region}
      kubeconfigBase64: ${beta-x86-64-aws-k8s-128.encodedKubeconfig}
      plugin: "e2e"
      mode: quick
  dependsOn: []
  resources: [beta-x86-64-aws-k8s-128-instances, beta-x86-64-aws-k8s-128]

Resource gets created, tests run to completion:

$ cli --kubeconfig etung-testsys.kubeconfig status
 NAME                                                 TYPE                        STATE                                         PASSED                    FAILED                   SKIPPED 
 beta-x86-64-aws-k8s-128-test                         Test                        passed                                             5                         0                      7386 
 beta-x86-64-aws-k8s-128                              Resource                    completed                                                                                                
 beta-x86-64-aws-k8s-128-instances                    Resource                    completed  

Terms of contribution:

By submitting this pull request, I agree that this contribution is dual-licensed under the terms of both the Apache License, version 2.0, and the MIT license.